Abstract
In machine learning, noise contained in the training dataset can be divided into attribute noise and label noise. Many works prove that label noise is more harmful compared to attribute noise. A set of noise filtering algorithms have been proposed to identify and remove noise prior to learning. However, almost all existing works solve this problem in a pure supervised way. It means noise identification is only based on the information of labeled data. In fact, unlabeled data are available in many applications, and the amount of unlabeled data is usually much bigger than labeled data. Therefore, in this paper, we consider to make use of unlabeled data to improve the performance of noise filtering. Tri-training is a powerful semi-supervised learning algorithm. It is adopted in this work because it is independent in the view of data. Finally, a set of experiments are conducted to prove the effectiveness of the proposed method.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Zhu, X., Wu, X.: Class noise vs. attribute noise: a quantitative study. Artif. Intell. Rev. 22(3), 177–210 (2004)
Quinlan, J.R.: Induction of decision trees. Mach. Learn. 1(1), 81–106 (1986)
Brodley, C.E., Friedl, M.A.: Identifying mislabeled training data. J. Artif. Intell. Res. 11, 131–167 (1999)
Gamberger, D., Lavrač, N., Džeroski, S.: Noise elimination in inductive concept learning: a case study in medical diagnosis. In: Arikawa, S., Sharma, A.K. (eds.) ALT 1996. LNCS, vol. 1160, pp. 199–212. Springer, Heidelberg (1996). https://doi.org/10.1007/3-540-61863-5_47
Gamberger, D., Lavrac, N., Dzeroski, S.: Noise detection and elimination in data preprocessing: experiments in medical domains. Appl. Artif. Intell. 14(2), 205–223 (2000)
Rico-Juan, J.R., Inesta, J.M.: Adaptive training set reduction for nearest neighbor classification. Neurocomputing 138, 316–324 (2014)
Calvo-Zaragoza, J., Valero-Mas, J.J., Rico-Juan, J.R.: Improving kNN multi-label classification in Prototype Selection scenarios using class proposals. Pattern Recogn. 48(5), 1608–1622 (2015)
Kanj, S., Abdallah, F., Denoeux, T., Tout, K.: Editing training data for multi-label classification with the k-nearest neighbor rule. Pattern Anal. Appl. 19(1), 145–161 (2015)
Roli, F.: Multiple classifier systems. In: Li, S.Z., Jain, A.K. (eds.) Encyclopedia of Biometrics, pp. 1142–1147. Springer, New York (2015). https://doi.org/10.1007/978-1-4899-7488-4
Wozniak, M., Grana, M., Corchado, E.: A survey of multiple classifier systems as hybrid systems. Inf. Fusion 16, 3–17 (2014)
Kuncheva, L.I., Rodriguez, J.J.: A weighted voting framework for classifiers ensembles. Knowl. Inf. Syst. 38(2), 259–275 (2014)
Sun, S.: Local within-class accuracies for weighting individual outputs in multiple classifier systems. Pattern Recogn. Lett. 31(2), 119–124 (2010)
Saez, J.A., Galar, M., Luengo, J., Herrera, F.: Tackling the problem of classification with noisy data using Multiple Classifier Systems: analysis of the performance and robustness. Inf. Sci. 247, 1–20 (2013)
Zhao, Y.C., Li, W.Z., Wu, J., Lu, S.L.: In: 2015 IEEE Conference on Computer Communications, pp. 26–30. IEEE, Hong Kong (2015)
Zhao, Y.C., Li, W.Z., Lu, S.L.: J. Netw. Comput. Appl. 74, 11–20 (2016)
Saez, J.A., Galar, M., Luengo, J., Herrera, F.: Analyzing the presence of noise in multi-class problems: alleviating its influence with the one-vs-one decomposition. Knowl. Inf. Syst. 38(1), 179–206 (2014)
Barandela, R., Valdovinos, R.M., Sanchez, J.S.: New applications of ensembles of classifiers. Pattern Anal. Appl. 6(3), 245–256 (2003)
Sanchez, J.S., Kuncheva, L.I.: Data reduction using classifier ensembles. In: ESANN, pp. 379–384 (2007)
Zhou, Z.H., Li, M.: Tri-training: exploiting unlabeled data using three classifiers. IEEE Trans. Knowl. Data Eng. 17(11), 1529–1541 (2005)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer Nature Switzerland AG
About this paper
Cite this paper
Zhu, H., Liu, J., Wan, M. (2018). Label Noise Detection Based on Tri-training. In: Sun, X., Pan, Z., Bertino, E. (eds) Cloud Computing and Security. ICCCS 2018. Lecture Notes in Computer Science(), vol 11063. Springer, Cham. https://doi.org/10.1007/978-3-030-00006-6_56
Download citation
DOI: https://doi.org/10.1007/978-3-030-00006-6_56
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-00005-9
Online ISBN: 978-3-030-00006-6
eBook Packages: Computer ScienceComputer Science (R0)